Phase Equilibrium Behavior of the Binary Systems CO2 + Nonadecane and CO2 + Soysolv and the Ternary System CO2 + Soysolv + Quaternary Ammonium Chloride Surfactant

J. Chem. Eng. Data 2003, 48, 1401-1406 1401

Mohamed M. Elbaccouch,† Valeriy I. Bondar,‡ Ruben G. Carbonell,* and Christine S. Grant Box 7905sChE Department, North Carolina State University, Raleigh, North Carolina 27695-7905

Liquid phase and molar volume data were measured for the binary system CO2 + soysolv at (298.15, 313.15, 323.15, 333.15, and 343.15) K and the ternary system CO2 + soysolv + quaternary ammonium chloride surfactant at (298.15, 313.15, and 333.15) K, where the composition of soysolv to the surfactant is 99:1 wt % and 80:20 wt % on a CO2-free basis. Data were collected stoichiometrically with a high- pressure Pyrex glass cell, where no sampling or chromatographic equipment is required. The accuracy of the experimental apparatus was tested with phase equilibrium measurements for the system CO2 + nonadecane at 313.15 K. A pressure-decay technique was used to calculate the mass of CO2 loaded into the equilibrium section of the apparatus, and its accuracy was verified with a blank nitrogen experiment. The generated data show that CO2 modified soysolv is an effective transport medium for the quaternary ammonium chloride surfactant.

Introduction Understanding the effects of gas solubility and swollen volume on the phase behavior of a mixture is important in various thermodynamic and industrial applications. The feasibility of using CO2, modified with a food grade oil cosolvent, as an extraction medium for particular compounds in chemical processes is of industrial interest. CO 2 Figure 1. Didecyldimethylammonium chloride. has been receiving industrial attention due to its solvating power, which can be altered dramatically by a slight change Table 1. Soysolv Chemical Composition4 1 in temperature or pressure. In addition, CO is a benign - 2 compound structure MW/g‚mol 1 wt % solvent, inexpensive, recyclable, nontoxic, and not flam- mable.2 Quaternary ammonium chloride surfactant (BTC methyl linoleate C19H34O2 294.5 62.4 methyl oleate C19H36O2 296.5 22.5 1010-80) is being used in the regeneration of a specific methyl palmitate C H O 270.5 8.5 packed ion exchange bed in aqueous media.3 It is desirable 17 34 2 methyl linolenate C19H32O2 292.5 3.2 to be able to recycle the surfactant and reduce its waste methyl stearate C19H38O2 298.5 3.0 volume by transporting it with CO2. The surfactant and methyl palmitoleate C17H32O2 268.4 0.3 CO2 are not miscible, but the addition of soysolv as a others na na 0.1 cosolvent to CO2 can provide the necessary solubility enhancement for removing the surfactant from the slurry oil, composed of mixed fatty acid methyl esters. The after the regeneration process. In this study, solubility data chemical composition of soysolv is given in Table 1. Its ‚ -1 for the binary system CO + soysolv and the ternary system molecular weight and boiling point are 292 g mol and 2 419 K, respectively.4 Soysolv is safe, nontoxic, and biode- CO2 + soysolv + quaternary ammonium chloride surfactant were measured. gradable. It is being used as a solvent in many cleaning The quaternary ammonium chloride surfactant contains applications, such as asphalt removal, grease and oil clean 80 wt % didecyldimethylammonium chloride, 12 wt % up, adhesive removal, welding clean up, and rubber 4 ethanol, and 8 wt % water. The ethanol and water are compound cleaning. + stabilizing agents for the didecyldimethylammonium chlo- Solubility data for the system CO2 soysolv have 5 ride to prevent precipitation. The chemical structure of received very little attention, but there are phase behavior + didecyldimethylammonium chloride is given in Figure 1. data for some of the components of soysolv CO2, such as + + + The quaternary ammonium chloride surfactant has been CO2 methyl linoleate, CO2 methyl oleate, and CO2 6-8 used industrially in disinfectant, sanitizer, and fungicidal methyl palmitate. Stoldt and Brunner measured the products for hard surfaces in hospitals and public institu- solubility of several soybean oil deodorizer distillates 9 tions.4 Also, it is an effective algaecide in swimming pools (SODDs), obtained from different plants, with CO2. SODD and industrial water treatment. Soysolv is 100% soybean does not exist in nature, but the refining process of soybean oil produces SODD, which generally comprises mixtures * To whom correspondence should be addressed. Telephone: (919) 515- of fatty acid, sterols, hydrocarbons, and vitamins. The 5118. Fax: (919) 515-583. E-mail: [email protected]. composition of SODD is similar to that of soysolv in the † Current address: Florida Solar Energy Center, 1679 Clearlake Road, sense that both contain fatty acids and ester chains. Cocoa, FL 32922. E-mail: [email protected]. ‡ Current address: RTI International, 3040 Cornwallis Road, Research We built a phase equilibrium unit capable of generating Triangle Park, NC 27709. E-mail: [email protected]. quantitative solubility and molar volume data for CO2 in 10.1021/je020201d CCC: $25.00 © 2003 American Chemical Society Published on Web 10/16/2003 1402 Journal of Chemical and Engineering Data, Vol. 48, No. 6, 2003

Figure 2. Schematic diagram of the phase equilbrium ap- Figure 3. Diagram of the visual cell and the specially designed paratus: visual cell section, everything to the right of V1; metal fitting. cell section, everything to the left of V1 and below V2. The Pyrex glass visual equilibrium cell is illustrated in 1 the liquid-phase mixture. As a preliminary step, we tested Figure 3. It has a height of 10 in. with a /2 in. o.d. body, and it tapers to a 3 in. height and a 1/ in. o.d. capillary our solubility measurement capability using a system for 4 neck. The cell is attached to 1/ in. stainless steel tubing which literature data were available at similar conditions 16 through a specially designed fitting to achieve the required to those proposed for the soysolv systems. The phase metal-to-glass pressure seal. The specially designed fitting behavior of CO2 + nonadecane was selected for comparison 1 is made from two parts. The first part is a /4 in. female because it was well studied and the published data could fitting located on the capillary neck; it is not removable 10 be used to check the accuracy of our data. We found that from the cell body (Figure 3). The second part is a our solubility and molar volume data agreed with previous 1 removable fitting containing a /16 in. female fitting from 1 5 measurements. one end and a /4 in. male fitting from the other end. A /8 In summary, we report the liquid-phase and molar in. glass circular flange is placed on the top of the capillary neck to sandwich the two parts of the fitting together. The volume data for the binary systems CO2 + nonadecane at 313.15 K and CO + soysolv at (298.15, 313.15, 323.15, visual cell is reliable to pressures of at least 10 MPa. The 2 content of the cell is stirred to achieve phase equilibrium 333.15, and 343.15) K. Also, we report the liquid-phase and using a steel ball manually actuated up and down in the molar volume data of the ternary system CO + soysolv + 2 cell by a magnet. quaternary ammonium chloride surfactant at (298.15, The pressure inside the visual cell section is measured 313.15, and 333.15) K, where the composition of soysolv to using a Heise digital pressure transducer (P1) (model 901- the surfactant is 99:1 wt % and 80:20 wt % on a CO2-free B) having a maximum working pressure of 41 MPa. The basis. The Appendix summarizes the method used to digital indicator has a resolution of 0.69 kPa and a rated calculate the mass of CO2 injected into the equilibrium cell accuracy of (6.9 kPa. The pressure inside the metal cell for the stoichiometric measurements. section is measured using an Omega pressure transducer (P2) (model DP25-S-A). Heise and Omega pressure trans- Experimental Section ducers are calibrated by their manufacturers using instru- mentation and standards that are traceable to the U.S. Apparatus. The data were generated using a stoichio- National Institute of Standards and Technology (NIST). metric technique, which involves an indirect determination Visual Cell Calibration. A calibration curve is pre- of equilibrium phase compositions. The technique does not pared relating the volume of a liquid (hexadecane, MW require sampling of equilibrium phases; it utilizes mass 226.45 g‚mol-1) in the visual cell to its height. The height balances and a volumetrically calibrated cell in determining of the liquid is measured using a cathetometer (Gaertner phase compositions. The setup used in this study is Scientific Co.; model 2778-A), which allows the volumes to represented schematically in Figure 2. It consists mainly be read in the visual cell to (0.005 cm3. The average of a visual cell section and a metal cell section. Everything deviation in the calibration curve is 0.2%.A1cm3 syringe with a 12 in. needle is used to inject a known mass of located to the right of valve 1 (V1) is considered part of the hexadecane into the visual cell (Mettler Toledo Inc.; model visual cell section. Everything located to the left of V1 and below V is considered part of the metal cell section. The AB-204). The mass of the liquid injected into the visual 2 ( visual and metal sections are housed in an acrylic water cell has an accuracy of 0.1 mg. From the known mass and density of hexadecane (0.773 g‚cm-3 at 22.4 °C), the bath. The temperature of the bath is controlled using a injected volume can be calculated, and the corresponding heater (Cole Parmer Instrument Co.; Model 12112-10) with height is measured. In this work, the average mass of liquid an accuracy of (0.1 °C. The temperature of the content of injected was within 0.0035 mg. The corresponding average the system is taken to be the temperature of the liquid liquid height and volume were within 0.6 cm and 2.69 cm3. bath, which is read directly from the heater’s temperature Assumption. We generated our data assuming that the 1 indicator. The /16 in. tubing located outside the water bath vapor phase in the visual cell is pure CO2. Generating and connecting the pressure indicator to the visual cell is multiphase equilibrium data using a stoichiometric tech- wrapped with a heating tape and heated to the same nique along with the assumption that the vapor phase temperature of the water bath to maintain a uniform is pure CO2 has been used extensively by Brennecke and temperature throughout the system. Valves V3 and V4 are co-workers, Khon and co-workers, and Luks and co- 10-16 16 used to apply a vacuum to the system and inject CO2, workers. Chen et al. generated multiphase equilib- respectively. rium data using the stoichiometric model for the systems Journal of Chemical and Engineering Data, Vol. 48, No. 6, 2003 1403

Table 2. Comparison between Vapor Pressures17 of n-Nonadecane and Methyl Oleate (mmHg) 298 K 313 K 323 K 333 K 343 K nonadecane 6.0 × 10-5 3.7 × 10-4 1.1 × 10-3 3.1 × 10-3 7.9 × 10-3 methyl 6.3 × 10-6 4.7 × 10-5 1.6 × 10-4 4.9 × 10-4 1.4 × 10-3 oleatea

a 22.5 wt % in soysolv. benzene + butane and benzene + propane at temperatures below 273 K. Foreman and Luks12 measured multiphase equilibrium data for the system n-tetradecane + CO2 in the presence of approximately 1 wt % of methanol at 303 K. In this study, since the composition of ethanol in the soysolv-surfactant-CO2 mixture is at most 2 wt %, the ethanol partition into the vapor phase is assumed to be negligible. Table 2 shows the vapor pressures of nonadecane and Figure 4. Pressure vs CO2 liquid-phase mole fraction for the + 10 methyl oleate at several temperatures.17 The data of Zou system CO2 nonadecane. and co-workers indicate that at 60 °C the vapor mole Table 3. Liquid-Phase Mole Fraction and Molar Volume fractions of methyl linoleate in CO2 and methyl oleate in of the Mixture as a Function of Pressure for the CO2 are 0.0103 at 91.0 bar and 0.0111 at 92.9 bar, Liquid-Vapor Isotherm of the CO2 (1) + n-Nonadecane respectively.7,8 Lockermann reports that the vapor mole System fraction of methyl palmitate in the system CO + methyl 3 -1 2 T/K P/bar x1 molar vol/cm ‚mol palmitate at 60 °C and 90 bar is 0.0006.6 The vapor 313.15 2.93 0.0379 337.9 pressure of nonadecane at 40 °C is similar to that of methyl 313.15 7.92 0.0999 319.1 oleate at 60 °C. So, we might expect their corresponding 313.15 14.20 0.174 296.3 vapor pressures to be similar. 313.15 21.24 0.247 274.3 With the above information as a background, we would 313.15 28.61 0.318 253.3 like to comment on the effect of the vapor-phase assump- 313.15 35.67 0.378 235.5 tion on our data. For each liquid-vapor isotherm, only one 313.15 44.52 0.448 214.6 313.15 53.95 0.513 196.0 liquid injection was utilized. This is an important approach 313.15 64.02 0.566 181.6 to avoid experimental errors which can result from multiple 313.15 74.56 0.631 162.6 injections of liquid. Since the height of the liquid mixture 313.15 80.34 0.674 148.1 increases by adding CO2, any error associated with the 313.15 86.38 0.723 129.9 vapor-phase assumption diminishes because the vapor headspace decreases. Injecting a large volume of liquid in Since we are assuming that the vapor space above the order to minimize the vapor headspace causes the meniscus liquid mixture in the visual cell section is pure CO2, the of the liquid mixture to reach the neck of the visual cell density of CO2 in the vapor space can be calculated from before all of the vapor-liquid isotherm is covered. So, there the temperature of the water bath and the Heise pressure are limits associated with the volume of the liquid injected indicator. By subtracting the volume of the liquid mixture into the visual cell. As was pointed out in the previous from the total volume of the visual cell section (calculated section, the average mass of liquid injected in this study in the Appendix), the volume of the vapor space above the 3 was within 2.69 cm . This is roughly 33% of the total liquid phase is obtained. From the density of CO2 in the volume of the visual cell (9.489 cm3; Appendix). Our vapor space and the volume of the vapor space, the mass estimation of the required liquid volume allows the collec- of CO2 in the vapor space is obtained. By subtracting the tion of all of the solubility data before the meniscus of the mass of CO2 in the vapor space from the total mass of CO2 liquid mixture reaches the neck of the visual cell, and it injected in the visual cell (calculated in the Appendix), the diminishes the error associated with the vapor-phase mass of CO2 in the liquid phase is obtained. The molar assumption. volume of the liquid phase mixture can be calculated be Chemicals. CO2 of minimum purity 99.99% was ob- dividing the volume of the liquid phase mixture by the total tained from National Specialty Gases. Hexadecane (99%) number of moles (CO2 and hydrocarbon) in the liquid and nonadecane (99.9%) were obtained from Aldrich Chemi- mixture. cal Co. Soysolv was obtained from Steyer Farms, Inc. The Liquid-phase and molar volume data for CO2 + nona- quaternary ammonium surfactant (BTC-1010-80) was sup- decane at 313.15 K are presented in Table 3. Figures 4 and plied by Sandia National Laboratories. 5 show that our data for the system CO2 + nonadecane Procedures. A known mass of the heavy hydrocarbon agree well with those reported by Luks and co-workers10 (i.e. nonadecane) was added to the visual cell using a 12 at 313.05 K. Table 4 and Figure 6 contain liquid-phase and in. needle syringe. The metal cell section and the vapor molar volume data for the system CO2 + soysolv at (298.15, space above the liquid in the visual cell section were 313.15, 323.15, 333.15, and 343.15) K. Tables 5 and 6 vacuumed, and CO2 was added to the metal cell section. present liquid-phase and molar volume data for the sys- Valve V1 was opened to transfer CO2 from the metal cell tems CO2 + 99:1 and 80:20 wt % soysolv-quaternary section to the visual cell section. The mass of CO2 inside ammonium surfactant at (298.15, 313.15, and 333.15) K. the metal cell section before opening V1 minus the mass of The corresponding solubility plots are given in Figures 7 CO2 inside the metal cell section after opening V1 was equal and 8. The presented data show that the liquid-phase to the mass of CO2 transferred to the visual cell section mixtures are stable and homogeneous under the operating (Appendix). The steel ball inside the visual cell was conditions (i.e. no surfactant precipitation or separate manually actuated using an external magnet to achieve phase formation). Figure 9 shows the pressure as a function equilibrium. of the mixture liquid-phase molar volume for the systems 1404 Journal of Chemical and Engineering Data, Vol. 48, No. 6, 2003

Table 5. Liquid-Phase Mole Fraction and Molar Volume of the Mixture as a Function of Pressure for Liquid-Vapor Isotherms of CO2 (1) + a 99:1 wt % Mixture of Soysolv + BTC-1010-80a TP mol vol TP mol vol

3 -1 3 -1 K bar x1 cm ‚mol K bar x1 cm ‚mol 298.15 2.01 0.0477 318.5 313.15 50.06 0.615 158.5 298.15 5.14 0.121 297.7 313.15 56.41 0.659 145.9 298.15 9.78 0.217 270.8 313.15 61.40 0.694 135.4 298.15 14.92 0.306 245.5 313.15 69.17 0.740 123.1 298.15 20.87 0.399 218.3 313.15 76.42 0.789 107.8 298.15 26.23 0.469 198.3 313.15 83.22 0.835 88.4 298.15 31.96 0.540 178.5 313.15 87.38 0.862 76.1 298.15 37.74 0.603 160.1 298.15 41.58 0.641 149.4 333.15 2.88 0.0478 328.9 298.15 43.79 0.661 143.6 333.15 8.29 0.123 307.3 298.15 47.13 0.691 135.1 333.15 16.28 0.228 275.7 Figure 5. Pressure vs mixture liquid-phase molar volume for the 298.15 49.59 0.718 127.1 333.15 25.15 0.323 247.7 10 333.15 34.60 0.408 222.3 system CO2 + nonadecane. 313.15 2.30 0.048 324.9 333.15 43.64 0.475 202.5 313.15 7.47 0.144 296.2 333.15 51.76 0.530 186.5 313.15 16.49 0.293 251.6 333.15 66.13 0.606 165.3 313.15 25.75 0.403 220.1 333.15 79.65 0.666 148.0 313.15 33.39 0.479 197.5 333.15 87.40 0.691 132.0 313.15 40.38 0.544 178.7

a BTC-1010-80 is didecyldimethylammonium chloride.

enhances the solubility of the CO2 in the liquid mixture.

Conclusions

A volumetrically calibrated visual cell was used to measure the solubility of CO2 in soysolv and in a soysolv + quaternary ammonium chloride surfactant mixture. The liquid-phase compositions are determined by stoichiometry Figure 6. Pressure vs CO2 liquid-phase mole fraction for the assuming that the vapor phase is pure CO2. The challenge system CO2 + soysolv. of accurately calibrating the total volume of the visual cell Table 4. Liquid-Phase Mole Fraction and Molar Volume and calculating the mass of CO2 has been overcome by of the Mixture as a Function of Pressure for the utilizing a dual-volume technique. The accuracy of our - + Liquid Vapor Isotherms of the CO2 (1) Soysolv System techniques is verified by successfully reproducing solubility TP mol vol TP mol vol data for the system CO2 + nonadecane. 3 -1 3 -1 K bar x1 cm ‚mol K bar x1 cm ‚mol The generated data demonstrate the capability of CO2 298.15 1.62 0.0557 311.4 323.15 2.468 0.045 327.3 298.15 5.16 0.134 288.2 323.15 8.129 0.140 297.6 modified soysolv to be an effective transport solvent for the 298.15 9.70 0.230 262.5 323.15 14.851 0.235 271.2 quaternary ammonium chloride surfactant. The CO2-BTC 298.15 14.40 0.312 239.8 323.15 24.035 0.338 241.9 surfactant mixture is not miscible, but soysolv is miscible 298.15 19.84 0.396 215.8 323.15 35.74 0.444 211.8 in the BTC surfactant. Figure 7 shows that CO is soluble 298.15 26.92 0.488 190.0 323.15 48.06 0.529 189.0 2 298.15 29.05 0.515 182.1 323.15 56.87 0.578 176.3 in soysolv under 100 bar. This work demonstrates a 298.15 33.17 0.563 168.8 323.15 80.27 0.706 143.2 heterogeneous separating process comprised of three steps.18 298.15 37.94 0.612 154.8 323.15 87.47 0.748 129.4 First, two immiscible phases are formed containing CO2 298.15 41.74 0.646 146.3 298.15 44.71 0.675 138.1 333.15 2.37 0.0387 322.0 in the first phase and the insoluble surfactant in the second 298.15 46.93 0.696 132.0 333.15 8.37 0.124 278.3 phase. Second, soysolv is added to the surfactant to produce 298.15 48.58 0.714 126.7 333.15 15.92 0.224 268.6 a completely soluble mixture in the second phase. Third, 333.15 26.52 0.337 236.6 CO is added to the second phase, in which CO becomes 313.15 2.06 0.044 326.3 333.15 36.34 0.422 212.6 2 2 313.15 6.01 0.116 305.4 333.15 46.79 0.499 190.9 soluble in the soysolv-surfactant mixture. Figures 8 and 313.15 8.91 0.166 290.5 333.15 57.03 0.563 172.9 9 show that CO2 is soluble in the soysolv-surfactant 313.15 13.82 0.239 269.3 333.15 70.89 0.632 154.1 mixture and as a result CO2 can effectively extract the BTC 313.15 20.25 0.321 245.2 333.15 82.98 0.686 139.0 313.15 27.84 0.404 220.2 surfactant. Also, the figures show that the solubility of CO2 313.15 35.44 0.477 200.1 343.15 2.23 0.0393 325.4 in the liquid phase of the mixture increases as the mass 313.15 43.26 0.544 181.1 343.15 8.73 0.120 303.2 percentage of the surfactant in soysolv increases. This is 313.15 49.72 0.598 165.7 343.15 20.54 0.245 266.8 indicated by the increase in the molar volume of the liquid 313.15 61.72 0.667 148.4 343.15 33.67 0.358 233.9 313.15 74.01 0.751 124.2 343.15 45.96 0.442 210.6 mixture as the mass of the BTC surfactant increases in 313.15 83.30 0.824 99.2 343.15 57.41 0.509 191.5 the mixture. Comparing the pressure ranges of Figures 313.15 87.66 0.862 83.1 343.15 70.06 0.568 175.0 7-9, the solubilities of the isothermal binary and ternary 313.15 90.17 0.877 76.0 343.15 84.78 0.628 159.5 mixtures are within the same range. CO2 + soysolv, CO2 + 99:1 wt % soysolv-BTC, and CO2 + 80:20 wt % soysolv-BTC at (298.15 and 333.15) K. The It should be noted that a second liquid-phase rich in CO2 figures show that the addition of the BTC surfactant was observed at 25 °C and 62 bar. Also, data were not Journal of Chemical and Engineering Data, Vol. 48, No. 6, 2003 1405

Table 6. Liquid-Phase Mole Fraction and Molar Volume of the Mixture as a Function of Pressure for Liquid-Vapor Isotherms of CO2 (1) + a 80:20 wt % Mixture of Soysolv + BTC-1010-80a TP mol vol TP mol vol

3 -1 3 -1 K bar x1 cm ‚mol K bar x1 cm ‚mol 298.15 1.88 0.0441 333.4 313.15 79.33 0.795 105.6 298.15 4.65 0.105 314.0 313.15 82.98 0.830 91.5 298.15 9.68 0.206 282.3 313.15 85.74 0.851 82.9 298.15 16.12 0.316 249.0 313.15 87.51 0.864 77.0 298.15 26.40 0.459 205.3 313.15 89.01 0.872 72.8 298.15 38.06 0.581 169.6 298.15 44.34 0.641 152.3 333.15 3.41 0.0499 340.9 298.15 48.10 0.681 140.1 333.15 8.96 0.119 217.0 298.15 50.68 0.704 133.5 333.15 15.42 0.193 292.6 333.15 22.61 0.263 270.6 313.15 3.48 0.0649 331.4 333.15 29.66 0.324 250.9 313.15 7.48 0.134 309.8 333.15 38.20 0.393 228.9 313.15 11.62 0.199 289.1 333.15 44.00 0.435 215.5 313.15 16.95 0.271 266.4 333.15 50.46 0.479 202.1 313.15 21.69 0.333 247.5 333.15 57.51 0.521 188.9 313.15 26.37 0.386 231.2 333.15 64.31 0.559 177.4 313.15 31.41 0.439 214.7 333.15 72.15 0.599 165.9 313.15 37.85 0.590 197.2 333.15 76.68 0.621 158.9 313.15 43.11 0.540 184.2 333.15 82.11 0.648 150.5 313.15 49.72 0.592 168.4 333.15 87.62 0.679 140.4 313.15 54.72 0.626 158.6 333.15 91.34 0.695 135.1 313.15 59.00 0.657 148.7 333.15 93.72 0.704 132.3

a BTC-1010-80 is didecyldimethylammonium chloride.

Figure 9. Pressure vs mixture liquid-phase molar volume for the systems CO2 + 99:1 and 80:20 wt % soysolv-BTC at (A) 298.15 K and (B) 333.15 K.

solubility data for multiple liquid phases at a larger range of temperatures and pressures.19,20

Acknowledgment We would like to thank Professors Joan Brennecke and James Kohn from the University of Notre Dame for their assistance in building the equilibrium visual cell. Figure 7. Pressure vs CO2 liquid-phase mole fraction for the + - system CO2 99:1 wt % soysolv BTC. Appendix Phase equilibrium experiments require precise and expensive equipment for measurements of pressure, temperature, and volume. Luks and co-workers10-13 and Kohn and co-workers14 used Ruska pumps (Ruska Instrument Co.) to calculate the mass of CO2 injected into the equilibrium cells. As an alternative technique, we calculated the mass of CO2 in the visual cell section by utilizing a dual volume system, which was considered by a number of investigators for the measurement of gas sorption in 21-25 polymers. The mass of CO2 injected into the visual cell section is equal to the difference in the mass of CO2 inside the metal cell section, calculated from the gas law, before and after opening valve V1. The following is a procedure for calculating the total volumes of the metal cell section and visual cell section. Figure 8. Pressure vs CO2 liquid-phase mole fraction for the system CO2 + 80:20 wt % soysolv-BTC. The experimental apparatus was vacuum pumped, and the water bath was heated to 40 °C. Valve V1 was closed generated beyond 90 bar due to the pressure limit of the and the visual cell section was pressurized with nitrogen. apparatus (100 bar). That motivates us to suggest the use Then, valve V1 was opened, letting nitrogen expand into of a different stoichiometric apparatus with a higher the metal cell section. Since the total mass of nitrogen pressure limit in order to have the capability to generate before and after opening V1 is the same, mass balance 1406 Journal of Chemical and Engineering Data, Vol. 48, No. 6, 2003 equations can be expressed as follows. (2) Canelas, D.; DeSimone, J. Polymerizations in Liquid and Super- critical Carbon Dioxide. Adv. Polym. Sci. 1997, 133, 103-140. ) (3) Wang, J.; Even, B. Personal Communication. Sandia National n(before expansion) n(after expansion) (1) Laboratories, Livermore, CA 94550. (4) Styer Farms, Inc. Technical Communications, Tiffin, OH 44883. + s PiVG Pf(VG VM) (5) Fogg, P. Solubility Data Series Carbon Dioxide in Non-Aqueous ) (2) Solvents at Pressures Less Than 200 Kpa; Pergamon Press: ZiRT ZfRT Oxford, 1992; Vol. 50, pp 444-445. (6) Lockermann, A. C. High-Pressure Phase Equilibria and Densities of the Binary Mixtures Carbon Dioxide-Oleic Acid, Carbon where n is the number of moles of nitrogen, and VG and Dioxide-Methyl Myristate, and Carbon Dioxide-Methyl Palmitate VM are the total volumes of the visual cell section and metal and the Ternary Mixture Carbon Dioxide-Methyl Myristate- - cell section, respectively. Pi is the initial pressure in the Methyl Palmitate. Chem. Eng. Process. 1994, 33, 171 187. visual cell section before expansion, P is the final pressure (7) Zou, M.; Yu, R. Z.; Rizvi, H. S.; Zollweg, A. J. Fluid-Liquid f Equilibria of Ternary Systems of Fatty Acids and Fatty Esters in the visual and metal cell sections after expansion, R is in Supercritical CO2. J. Supercrit. Fluids 1990, 3,85-90. the universal gas constant, and T is the temperature (40 (8) Zou, M.; Yu, R. Z.; Kashulines, S. S.; Rizvi, H. S.; Zollweg, A. J. Fluid-Liquid-Phase Equilibria of Fatty Acids and Fatty Acid °C). Zi and Zf are the compressibility factors of nitrogen at 26 Methyl Esters in Supercritical CO2. J. Supercrit. Fluids 1990, 3, pressures Pi and Pf, respectively. The volume ratio of the 23-28. metal and visual sections can be expressed by arranging (9) Stoldt, J.; Brunner, G. Phase Equilibrium Measurements in eq2as Complex Systems of Fats, Fat Compounds in Supercritical Carbon Dioxide. Fluid Phase Equilib. 1998, 146, 269-295. (10) Fall, D.; Fall, J.; Luks, K. Liquid-Liquid-Vapor Immiscibility V P Z + M ) i f - ) limits in Carbon Dioxide n-Paraffin Mixtures. J. Chem. Eng. 1 Q1 (3) Data 1985, 30,82-88. VG PfZi (11) Fall, D.; Luks, K. Phase Equilibria Behavior of the Systems Carbon Dioxide + n-Dotriacontane and Carbon Dioxide + n- Docosane. J. Chem. Eng. Data 1984, 29, 413-417. The constant Q1 was determined on the basis of 10 parallel (12) Foreman, C.; Luks, K. Partial Miscibility Behavior of the Ternary expansion experiments at 40 °C and different pressures Mixture Carbon Dioxide + n-Tetradecane + Methanol. J. Chem. while placing the metal cell section under a vacuum prior Eng. Data 2000, 45, 334-337. to each pressure load. The average Q1 value was equal to (13) Lam, D.; Luks, K. Multiphase Equilibrium Behavior of the Mixture Ethane + Methanol + 1-Decanol. J. Chem. Eng. Data 0.7658 with a 0.006 standard deviation. 1991, 36, 307-311. A similar set of expansion measurements was conducted (14) Tarantino, D.; Kohn, J.; Brennecke, J. Phase Equilibrium Be- with stainless steel spheres of a known volume V inside havior of the Carbon Dioxide + Benzophenone Binary System. B - the metal cell. The volume of the spheres was determined J. Chem. Eng. Data 1994, 39, 158 160. (15) Tiffin, D.; Luks, K. Solid Hydrocarbon Solubility in Liquid 3 by a picnometric technique and was found to be 2.249 cm Methane-Ethane Mixtures along Three-Phase Solid-Liquid- with a 0.018 standard deviation. The calibration procedure Vapor Loci. J. Chem. Eng. Data 1979, 24, 306-310. was the same as that described above, and eq 2 can be (16) Chen, W. L.; Luks, K.; Kohn, J. Three-Phase Solid-Liquid-Vapor Equilibria of the Binary Hydrocarbon Systems Propane-Benzene, expressed as Propane Cyclohexane, n-Butane-Benzene, n-Butane-Cyclohexane, n-Butane-n-Decane, and n-Butane-n-Dodecane. J. Chem. Eng. / V V - V P Z Data 1981, 26, 310-312. M ) M B ) i f - 1 ) Q (4) (17) Yaws, L. C. Chemical Properties Handbook; McGraw-Hill: New V V P Z 2 York, 1999. G G f i (18) Romack, T. J.; McClain, J. B.; Stewart, G. M.; Givens, R. D. / Carbon Dioxide Cleaning and Separation Systems. Patent US- where VM is the volume of the metal cell section minus 6120613, 2000. the volume of the spheres. The constant Q was found to (19) Khan, V. A. Solubility of Gases in Molten Polymers at High 2 Pressures. Dissertation, Wayne State University, 1998. be equal to 0.5843 with a 0.003 standard deviation. The (20) Stradi, G. B. Measurement and Modeling of the Phase Behavior constant Q2 was based on over 20 different pressure loads of High-Pressure-Reaction Mixtures and the Computation of of nitrogen at 40 °C, while placing the metal cell section Mixture Critical Points. Dissertation, University of Notre Dame, 2000. under a vacuum prior to each expansion. The final expres- (21) Bondar, V.; Freeman, B. Sorption of Gases and Vapors in an sions for the visual cell section and the metal cell section Amorphous Glassy Perfluorodioxole Copolymer. Macromolecules are given as follows. 1999, 32, 6163-6171. (22) Koros, J. W.; Paul, R. D. Design Considerations for Measurements of Gas Sorption in Polymers by Pressure Decay. J. Polym. Sci. VB 1976, 14, 1903-1907. V ) (5) G Q - Q (23) Sada, E.; Kumazawa, H.; Xu, P. Sorption and diffusion of carbon 1 2 dioxide in polyimide films. J. Appl. Polym. Sci. 1988, 35, 1497- 1509. ) VM VGQ1 (6) (24) Sanders, E. S.; Koros, W. J.; Hopfenberg, H. B.; Stannett, V. T. Pure and mixed gas sorption of carbon dioxide and ethylene in poly(methyl methacrylate). J. Membr. Sci. 1984, 18,53-74. The volumes of the visual cell section VG and the metal (25) Vieth, R. W.; Tam, M. P.; Michaels, S. A. Dual Sorption Mecha- 3 3 cell section VM were found to be 9.489 cm and 12.391 cm , nisms in Glassy Polystyrene. J. Colloid Interface Sci. 1966, 22, respectively. A blank nitrogen experiment based on over 360-370. (26) Ely, J. F.; Magee, J. W.; Haynes, W. M. Thermophysical Properties 50 runs was conducted to verify the accuracy of the volumes for Special High CO2 Content Mixtures; U.S. National Institute VG and VM. The absolute percentage error between the of Standards and Technology (NIST) Research Report RR-110; moles of nitrogen removed from the metal cell section and Gas Processors Association: Tulsa, OK, p 987. the corresponding moles transferred to the visual cell section was within 0.28%. Received for review November 18, 2002. Accepted July 11, 2003. This material is based upon work supported by the STC Program of Literature Cited the National Science Foundation under Agreement No. CHE- 9876674. (1) McHugh, M.; Krukonis, V. Supercritical Fluid Extraction, 2nd ed.; Butterworth-Heinemann: Boston, MA, 1994. JE020201D